DocumentCode :
1622750
Title :
A deformable model for hippocampus segmentation: improvements and extension to 3D
Author :
Ghanei, Amir ; Soltanian-Zadeh, Hamid ; Windham, Joe P.
Author_Institution :
Henry Ford Health Syst., Detroit, MI, USA
Volume :
3
fYear :
1996
Firstpage :
1797
Abstract :
The application of a deformable model to the segmentation of hippocampus in brain MRI has been investigated. Common problems of the model in this case and similar cases have been discussed and solved. A new method for extracting discontinuous boundaries of an object with multiple unwanted edges has been developed. This method is based on detecting and following the edge by external forces. For improving the contour stability, its movement has been limited. Also, adaptive values for internal force weights have been used. In the next step, the model has been extended to 3D which is a Deformable Surface Model. A geometric structure used for this purpose. This helps in definition of normal vectors and internal forces. Finally, a method for generating the initial volume from individual initial polygons has been developed
Keywords :
biomedical NMR; brain; brain models; edge detection; image segmentation; medical image processing; 3D extension; adaptive values; brain MRI; contour stability; deformable model; deformable surface model; discontinuous boundaries extraction; geometric structure; hippocampus segmentation; initial polygons; internal force weights; internal forces; magnetic resonance imaging; medical diagnostic imaging; multiple unwanted edges; normal vectors; Biomedical imaging; Brain; Deformable models; Geometry; Hippocampus; Image edge detection; Image segmentation; Magnetic resonance imaging; Solid modeling; Spline;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Nuclear Science Symposium, 1996. Conference Record., 1996 IEEE
Conference_Location :
Anaheim, CA
ISSN :
1082-3654
Print_ISBN :
0-7803-3534-1
Type :
conf
DOI :
10.1109/NSSMIC.1996.587978
Filename :
587978
Link To Document :
بازگشت